Laser gyroscopes have been proposed in which waves travel in opposite directions through a laser medium so that rotation of the laser medium about an axis will produce a difference frequency. However, unless the frequencies of the waves are spaced a substantial distance apart, the coupling of the wave propagating in one direction with the wave propagating in the opposite direction in the laser material can produce a combined laser action pulling the two frequencies toward each other and producing a condition known as lock-in.
Lock-in limits the use of laser gyroscopes since at low rotation rates where lock-in normally occurs, the output drops to zero, producing a range of rotation rates where there is no output since the clockwise and counterclockwise waves have the same frequency. Coupling between the waves may occur in many ways, including back scattering of energy from elements of the laser system such as window interfaces or other transitions from one medium into another.
Another possible source of coupling within the laser medium itself occurs when two waves travelling in opposite directions are in an instantaneous phase relationship where they compete for gain from atoms having a low velocity in the direction of propagation of the waves. The probablity of lock-in which affects the width of the inaccurate zero output region of the gyro in general will increase as the loop gain of the laser increases.
If the frequencies are spaced apart a substantial distance, for example by the use of devices that produce different delays in one direction than in the other, this frequency difference must be accurately maintained. Attempts to achieve an accurate frequency separation by switching a Faraday rotator from one condition to another have proved impractical since the accuracy of the AC switching waveform must be perfectly symmetrical to a degree substantilly greater than one part in 10.sup.6.
In addition, if a non-switched Faraday rotator were used to produce the different frequencies for opposite propagation directions, variations in the Faraday rotator could produce frequency variations greater than the dyroscopic rotational frequency changes, hence rendering the system inaccurate.